QCloud2: An Improved Cloud-based Quality-Control System for Mass-Spectrometry-based Proteomics Laboratories.

QCloud is a cloud-based system to support proteomics laboratories in daily quality assessment using a user-friendly interface, easy setup, and automated data processing. Since its release, QCloud has facilitated automated quality control for proteomics experiments in many laboratories. QCloud provides a quick and effortless evaluation of instrument performance that helps to overcome many analytical challenges derived from clinical and translational research. Here we present an improved version of the system, QCloud2. This new version includes enhancements in the scalability and reproducibility of the quality-control pipelines, and it features an improved front end for data visualization, user management, and chart annotation. The QCloud2 system also includes programmatic access and a standalone local version.

Can bioinformatics help in the identification of moonlighting proteins?

Protein multitasking or moonlighting is the capability of certain proteins to execute two or more unique biological functions. This ability to perform moonlighting functions helps us to understand one of the ways used by cells to perform many complex functions with a limited number of genes. Usually, moonlighting proteins are revealed experimentally by serendipity, and the proteins described probably represent just the tip of the iceberg. It would be helpful if bioinformatics could predict protein multifunctionality, especially because of the large amounts of sequences coming from genome projects. In the present article, we describe several approaches that use sequences, structures, interactomics and current bioinformatics algorithms and programs to try to overcome this problem. The sequence analysis has been performed: (i) by remote homology searches using PSI-BLAST, (ii) by the detection of functional motifs, and (iii) by the co-evolutionary relationship between amino acids. Programs designed to identify functional motifs/domains are basically oriented to detect the main function, but usually fail in the detection of secondary ones. Remote homology searches such as PSI-BLAST seem to be more versatile in this task, and it is a good complement for the information obtained from protein-protein interaction (PPI) databases. Structural information and mutation correlation analysis can help us to map the functional sites. Mutation correlation analysis can be used only in very restricted situations, but can suggest how the evolutionary process of the acquisition of the second function took place.

A novel subfamily of mouse cytosolic carboxypeptidases.

Nna1 is a recently described gene product that has sequence similarity with metallocarboxypeptidases. In the present study, five additional Nna1-like genes were identified in the mouse genome and named cytosolic carboxypeptidase (CCP) 2 through 6. Modeling suggests that the carboxypeptidase domain folds into a structure that resembles metallocarboxypeptidases of the M14 family, with all necessary residues for catalytic activity and broad substrate specificity. All CCPs are abundant in testis and also expressed in brain, pituitary, eye, and other mouse tissues. In brain, Nna1/CCP1, CCP5, and CCP6 are broadly distributed, whereas CCP2 and 3 exhibit restricted patterns of expression. Nna1/CCP1, CCP2, CCP5, and CCP6 were found to exhibit a cytosolic distribution, with a slight accumulation of CCP5 in the nucleus. Based on the above results, we hypothesized that Nna1/CCP1 and CCP2-6 function in the processing of cytosolic proteins such as alpha-tubulin, which is known to be modified by the removal of a C-terminal tyrosine. Analysis of the forms of alpha tubulin in the olfactory bulb of mice lacking Nna1/CCP1 showed the absence of the detyrosinylated form in the mitral cells. Taken together, these results are consistent with a role for Nna1/CCP1 and the related CCPs in the processing of tubulin.

Nna1-like proteins are active metallocarboxypeptidases of a new and diverse M14 subfamily.

Nna1 has some sequence similarity to metallocarboxypeptidases, but the biochemical characterization of Nna1 has not previously been reported. In this work we performed a detailed genomic scan and found >100 Nna1 homologues in bacteria, Protista, and Animalia, including several paralogs in most eukaryotic species. Phylogenetic analysis of the Nna1-like sequences demonstrates a major divergence between Nna1-like peptidases and the previously known metallocarboxypeptidases subfamilies: M14A, M14B, and M14C. Conformational modeling of representative Nna1-like proteins from a variety of species indicates an unusually open active site, a property that might facilitate its action on a wide variety of peptide and protein substrates. To test this, we expressed a recombinant form of one of the Nna1-like peptidases from Caenorhabditis elegans and demonstrated that this protein is a fully functional metallocarboxypeptidase that cleaves a range of C-terminal amino acids from synthetic peptides. The enzymatic activity is activated by ATP/ADP and salt-inactivated, and is preferentially inhibited by Z-Glu-Tyr dipeptide, which is without precedent in metallocarboxypeptidases and resembles tubulin carboxypeptidase functioning; this hypothesis is strongly reinforced by the results depicted in Kalinina et al. published as accompanying paper in this journal. Our findings demonstrate that the M14 family of metallocarboxypeptidases is more complex and diverse than expected, and that Nna1-like peptidases are functional variants of such enzymes, representing a novel subfamily (we propose the name M14D) that contributes substantially to such diversity.

TrSDB: a proteome database of transcription factors.

TrSDB-TranScout Database-(http://ibb.uab.es/trsdb) is a proteome database of eukaryotic transcription factors based upon predicted motifs by TranScout and data sources such as InterPro and Gene Ontology Annotation. Nine eukaryotic proteomes are included in the current version. Extensive and diverse information for each database entry, different analyses considering TranScout classification and similarity relationships are offered for research on transcription factors or gene expression.

Bioinformatics and Moonlighting Proteins.

Multitasking or moonlighting is the capability of some proteins to execute two or more biochemical functions. Usually, moonlighting proteins are experimentally revealed by serendipity. For this reason, it would be helpful that Bioinformatics could predict this multifunctionality, especially because of the large amounts of sequences from genome projects. In the present work, we analyze and describe several approaches that use sequences, structures, interactomics, and current bioinformatics algorithms and programs to try to overcome this problem. Among these approaches are (a) remote homology searches using Psi-Blast, (b) detection of functional motifs and domains, (c) analysis of data from protein-protein interaction databases (PPIs), (d) match the query protein sequence to 3D databases (i.e., algorithms as PISITE), and (e) mutation correlation analysis between amino acids by algorithms as MISTIC. Programs designed to identify functional motif/domains detect mainly the canonical function but usually fail in the detection of the moonlighting one, Pfam and ProDom being the best methods. Remote homology search by Psi-Blast combined with data from interactomics databases (PPIs) has the best performance. Structural information and mutation correlation analysis can help us to map the functional sites. Mutation correlation analysis can only be used in very specific situations - it requires the existence of multialigned family protein sequences - but can suggest how the evolutionary process of second function acquisition took place. The multitasking protein database MultitaskProtDB (http://wallace.uab.es/multitask/), previously published by our group, has been used as a benchmark for the all of the analyses.

Classification of common functional loops of kinase super-families.

A structural classification of loops has been obtained from a set of 141 protein structures classified as kinases. A total of 1813 loops was classified into 133 subclasses (9 betabeta(links), 15 betabeta(hairpins), 31 alpha-alpha, 46 alpha-beta and 32 beta-alpha). Functional information and specific features relating subclasses and function were included in the classification. Functional loops such as the P-loop (shared by different folds) or the Gly-rich-loop, among others, were classified into structural motifs. As a result, a common mechanism of catalysis and substrate binding was proved for most kinases. Additionally, the multiple-alignment of loop sequences made within each subclass was shown to be useful for comparative modeling of kinase loops. The classification is summarized in a kinase loop database located at http://sbi.imim.es/archki.

Including Functional Annotations and Extending the Collection of Structural Classifications of Protein Loops (ArchDB).

Loops represent an important part of protein structures. The study of loop is critical for two main reasons: First, loops are often involved in protein function, stability and folding. Second, despite improvements in experimental and computational structure prediction methods, modeling the conformation of loops remains problematic. Here, we present a structural classification of loops, ArchDB, a mine of information with application in both mentioned fields: loop structure prediction and function prediction. ArchDB (http://sbi.imim.es/archdb) is a database of classified protein loop motifs. The current database provides four different classification sets tailored for different purposes. ArchDB-40, a loop classification derived from SCOP40, well suited for modeling common loop motifs. Since features relevant to loop structure or function can be more easily determined on well-populated clusters, we have developed ArchDB-95, a loop classification derived from SCOP95. This new classification set shows a ~40% increase in the number of subclasses, and a large 7-fold increase in the number of putative structure/function-related subclasses. We also present ArchDB-EC, a classification of loop motifs from enzymes, and ArchDB-KI, a manually annotated classification of loop motifs from kinases. Information about ligand contacts and PDB sites has been included in all classification sets. Improvements in our classification scheme are described, as well as several new database features, such as the ability to query by conserved annotations, sequence similarity, or uploading 3D coordinates of a protein. The lengths of classified loops range between 0 and 36 residues long. ArchDB offers an exhaustive sampling of loop structures. Functional information about loops and links with related biological databases are also provided. All this information and the possibility to browse/query the database through a web-server outline an useful tool with application in the comparative study of loops, the analysis of loops involved in protein function and to obtain templates for loop modeling.